Construction of a 3D city map using EPI analysis and DP matching

In this paper, we propose an efficient method to make a 3D city map from real-world video data using digital map. To achieve this purpose, mainly two problems mentioned below exist. 1. Video data is usually huge, so suitable data structure is needed for an efficient handling 2. To construct 3D map from video data using digital map, matching between video data and digital map is needed To solve the first problem, we propose an automatic organization method focusing on video objects to describe video data. Actually, we divide spatio-temporal space into video objects which represent individual buildings by EPI analysis. To solve the second problem, we use DP matching method. When we apply DP matching method, we use depth information obtained from the real-world video by EPI analysis. And this method successfully improved the precision and reliability of the DP matching. On its implementation, we made two sample applications to demonstrate an effectiveness of our proposed method. First application is a “Interactive Q&A System” which extracts an video object from video data automatically and makes DPmatching between video data and digital map. Second application is a “3D modeling program” which makes an VRML 3D map from DPmatching result. Both applications prove that our method is working enough efficient for our purpose.